UCL Professor: Why AI’s Biggest Problem Might Be Energy, Not Intelligence

What happens when the future of AI depends less on bigger models and more on rebuilding the hardware underneath them?In this conversation, Anthony Kenyon explores semiconductors, nanoelectronics, and neuromorphic computing, and explains why the next leap in AI may come not from scaling the same architecture forever, but from designing systems that work more like the brain.But this isn’t just a story about faster chips.It’s about a deeper shift:🔹 What happens when today’s computing model becomes too energy-hungry to keep scaling?🔹 Could neuromorphic systems unlock a new generation of AI by copying the brain’s efficiency instead of brute-forcing more compute?🔹 And if that shift does happen, which countries, companies, and technologies will shape what comes next?At the centre of the discussion is a harder question:If current AI infrastructure is already running into limits around power, memory, and scale, can the next era of computing arrive before those bottlenecks slow everything down?Because the next breakthrough in AI may not come from bigger models, but from a completely different machine.